Making Knowledge - Life Out of Sequence - page 48

this sort of biological work—the emphasis falls on general problems, on
big patterns and correlations. The work of the Broad team becomes to
analyze and make sense of a vast output of data: comparing genomes,
describing the whole human microbiome, analyzing and classifying all
forms of cancer, mapping the totality of human haplotypes, and so on. 24
One informant vividly described this kind of task as creating a “Google
Earth for looking at the data.”
Whereas pre-informatic biologists usually dealt with a single organ-
ism, or perhaps even a single gene—devoting their careers and the careers
of their graduate students to characterizing it in detail—bioinformatic
biologists ask (and answer) bigger questions—questions that rely on
knowledge of all human genes, or a comparison across twenty or more
organisms. In short, they depend on having extraordinarily large vol-
umes of reliable and accessible data. Indeed, the biologists at the Broad
are in constant need of more sequence data. Although it is true that more
traditional biologists are also interested in general questions about life,
their day-to-day focus usually remains specifi c; without genome-scale
data (and bioinformatic techniques to digest them), answers to such
questions must rely on speculative extrapolation from a few examples.
The big questions remain the same, but the answers that can be given
are of a markedly different kind.
If molecular biology was about attention to detail and particularity,
bioinformatics is about generality and totality. Whereas a molecular bi-
ologist might spend a career working to characterize a particular gene,
asking questions about its function, its regulation, and its interactions,
the bioinformatician is interested in all genes, or all proteins, or all inter-
actions. Using computers means that it is possible to ask new, broader
kinds of questions. But working at this level requires a surrendering of
certain kinds of particular, granular knowledge. Using bioinformatics to
determine that 45% of genes in humans are regulated by two or more
different regulatory proteins tells the investigator nothing about which
particular proteins regulate a particular gene. Bioinformatics entails a
decision that certain types of knowledge are not important—that a cer-
tain amount of ignorance of detail is required. Biologists' response to a
growing amount of biological data was to fi nd a different route to turn
information into knowledge—there were too many data to continue to
use and process them in the way that molecular biology had done since
the 1950s. Instead, biologists began to use computers to manage what
could not be known about the details of particular genes, proteins, and
so on.
Paying attention to what bioinformatics does with data shows how